Garlef

13 ms·

211.
by Garlef 1y ago
Funny sidenote: The mobile client for "Magic: the Gathering - Arena" does not even support "roaming" for your phones connection: If you for example turn off wifi during a match, it looses (and can not reestablish) the fu
212.
by Garlef 1y ago
Maybe. But my experience as a mathematician tells me another part of that story. Certain fields are much more used to consuming (and producing) visual noise in their notation! Some fields have even superfluous parts in their definitions and
213.
by Garlef 1y ago
> "We have to be more efficient as we scale up" ... google is a 27 year old company ... google is one of the most valuable and influential companies currently operating i don't think the phrase "scaling up" makes
214.
by Garlef 1y ago
There's a fun experiment with toddlers where they re-enter a room but the car they just sat in was replaced by a tiny version: They will try to get into the car even though only their foot fits in. So size/scale is not as easy a c
215.
by Garlef 1y ago
I remember from the early days of midjourney (and SD,etc) that there was a point in time where midjourney had created more images than the whole of humanity had created in its entire existence up to that point (or something along the lines)
216.
by Garlef 1y ago
Tell me you don't have kids without telling me you don't have kids.
217.
by Garlef 1y ago
Maybe we can get a counter-AI that does the AI interview for us?
218.
by Garlef 1y ago
Is there an option to filter the list based on the measurements? I.e "context window > X, intelligence > Y, price < Z"? I'd love that. It seems the only filter options available are unrelated to the measured metrics.
219.
by Garlef 1y ago
The real rickrolling is to get everyone here on HN to read the lyrics.
220.
by Garlef 1y ago
I think there's a lot of room for even better UI. I'd love to have something that operates more at the codebase level. Autocomplete is very local. (Maybe "tab completion" when setting up a new package in a monorepo? Or m
221.
by Garlef 1y ago
One nice realization I had when using a similar feature in roo: You don't need a full agent library to write LLM workflows. Rather: A general purpose agent with a custom addition to the system prompt can be instructed to call other suc
222.
by Garlef 1y ago
Maybe enshitification is entering its dadaist phase? That's the best explanation I have for how youtube decided to implement this feature.
223.
by Garlef 1y ago
The real issue is not making the choice itself but rather bringing the knowledge about what the decision space is in the first place. (But sometimes the "knowledge" is also a limiting form of preconception)
224.
by Garlef 1y ago
Yes. But there's law firms who streamline such individual processes if the business case is actually large enough. For example there's a law that says the airline needs to pay you 400€(?) if your flight is delayed by more than 2h
225.
by Garlef 1y ago
Is selection really the issue? You'd still need to figure out what payload to give to the tool based on your context. But I guess depending on your business case it might be worth it. It's not something I'd do from the beginn
226.
by Garlef 1y ago
What you said is not correct! In detail: * "A list is not a monad" - True! * "List is a monad" - True. But I think "`List` is a monad" might be clearer. * "A list is an algebra for the `List` monad."
227.
by Garlef 1y ago
My last contact with AWS support (100€/month tier) was someone feeding me LLM generated slop that contained hallucinations about nonexistent features and configuration options.
228.
by Garlef 1y ago
> Having a conversation with a computer to generate code that I don't entirely understand and then have to try to review is just not fun. Same for me. But maybe that's ultimately an UX issue? And maybe things will straighten ou
229.
by Garlef 1y ago
Doesn't this mean: There's room for disruption/land grab? If the big corporations can't move fast enough and 100 startups gamble on getting there, eventually one of them will be successful.
230.
by Garlef 1y ago
No money mentioned? Empty talk.
231.
by Garlef 1y ago
You don't understand how that works: EU companies are required to act in compliance with the GDPR. This includes all sensitive data that is transfered to business partners. They must make sure that all partners handle the (sensitive pa
232.
by Garlef 1y ago
I think there's a world where you would use the code structuring of RSCs to compile a static page that's broken down into small chunks of html, css, js. Basically: If you replace the "$1" placeholders from the article wi
233.
by Garlef 1y ago
Maybe that's because we're conditioned by the UX of search. But another thing I find even more surprising is that, at least initially, many expected that the LLMs would give them access to some form of higher truth.
234.
by Garlef 1y ago
Yes . But maybe there's multiple such "conscious attention" instances at the same time. And "you" are only one of them.
235.
by Garlef 1y ago
I think Double Categories [1] would be a more appropriate setting: In a double category, the vertical and horizontal arrows are of different types. In usual commutative diagrams, they are of the same type. [1]: https://ncatlab.or
236.
by Garlef 1y ago
This made me happy: It offers an interesting take on AI. (After reflecting a bit on this I think this is for the following reason: Not only does this take a step back to offer a meta perspective. It also does so without falling into the tra
237.
by Garlef 1y ago
Have you considered that this might not be due to the model itself but due to less focus/time/money spent on alignment during the training? My guess is that this is a bit of a throwaway experiment before they actually spend millio
238.
by Garlef 1y ago
Because this approach also works for stuff that is not a class instance.
239.
by Garlef 1y ago
Support: Yes. But the UX is not optimized for this. Imagine trying to find a specific output/input that was good in the conversation tree.
240.
by Garlef 1y ago
Drinking has been an integral part of universities for a long, long time.

More ›